A New Approach to Arms Control

By Henry Kissinger

To halt the strategic arms race, the former Secretary of State recommends a dramatic new approach. His plan would scrap all MIR Vs starting in, say, 1990 and rely instead on a mix of mobile, single-warhead missiles. The U.S. would implement this plan even if the Soviets refused to go along

The nomination of a new director of the Arms Control and Disarmament Agency has prompted a major Senate debate over whether the Administration is seriously committed to arms control. The controversy misses the real question: To what kind of arms control should the Administration commit itself?

So far, the controversy has focused on negotiations over deployment of medium-range missiles in Europe. But even a success in these negotiations--likely now that the German elections are over--will make only a marginal contribution to the stability of the U.S.-Soviet strategic nuclear relationship. And here we are trapped in a conceptual crisis. For too long arms control and strategy have been proceeding on separate, increasingly incompatible tracks. Technology has driven weapons procurement at the same time that it has made irrelevant the traditional doctrines of arms control. Weapons systems and arms control schemes developed in isolation from each other had to be squeezed into a more or less arbitrary framework.

This article seeks to sketch an approach by which strategy and arms control can be reconciled and strategic stability achieved by ending the disproportion between warheads and launchers that is at the heart of the current strategic instability.

THE ORIGINS OF ARMS CONTROL

The concept of arms control evolved when a growing Soviet nuclear arsenal suddenly threw into doubt the comfortable premises of the decade after World War II. It had been complacently assumed that by means of the "balance of terror," technology supplied a shortcut to security. Even after we had lost our atomic monopoly, our superiority was so crushing that in 1954 Secretary of State John Foster Dulles could still declare a policy of massive retaliation, countering Soviet aggression anywhere by the threat to devastate the Soviet Union. There was little incentive to decide a question that seemed then only esoteric: What were rational targets for the apocalyptic arsenal we were assembling?

Soviet hydrogen weapons and Sputnik foreshadowed a kind of stalemate. Once general nuclear war threatened both sides with tens of millions of casualties, the very existence of nuclear arsenals came to be perceived by many as a menace. Traditional wars had been sustained by the conviction that the consequences of defeat or surrender were worse than the costs of resistance. The nuclear specter banished that conviction. Fewer and fewer objectives seemed worth the cost or the risk.

One result has been the growing determination to stabilize and ideally to reverse the arms race by negotiated agreement. There are at least two unprecedented aspects to the nuclear arms race. The destructiveness of the weaponry sets an upper limit beyond which additions to destructiveness become more and more marginal. At the same time, the complex technology of the nuclear age raises the danger of an automaticity that might elude rational control. For if one side should destroy the retaliatory force of its adversary, it would be in a position to impose its terms. That prospect could tempt the intended victim to undertake a "preemptive" first strike--or launch its weapons on warning. Mutual fear could turn a crisis into a catastrophe. Proponents of arms control thus saw it as their immediate objective to reduce the incentives and possibilities of surprise attack.

The goal was to reduce the vulnerability of strategic forces by maintaining symmetrical numbers of strategic weapons. If neither side could hope to destroy its opponent, the incentive for surprise attack would disappear in the face of certain and intolerable retribution. So long as missiles had single warheads and airplanes needed hours to reach their targets, a surprise attack would require a vast numerical preponderance. (Even with highly accurate missiles, an attacker probably would not risk a first strike without two warheads for each target, to allow for malfunctioning, hardened targets or errors in accuracy.)

In these circumstances, numerically equal retaliatory forces were rightly conceived as adequate insurance against surprise attack. The optimum total should be large enough so that it could be overwhelmed only by a violation of the agreement too large to be hidden. Yet it should involve a ceiling that would stop the accumulation of strategic weapons. This was the intellectual basis of the arms limitation talks proposed by President Johnson and implemented by President Nixon.

During 1969 and 1970, the Nixon Administration undertook painstaking studies to determine the lowest level above which a strategically significant violation could not be concealed. The culmination was the SALT agreements of 1972. These accords severely limited antiballistic missile defenses to discourage an aggressor from believing he could launch a surprise attack and then defend against a counterblow. The agreements also froze the number of offensive missiles for five years. At that point the Soviets had a numerical edge in missiles--though not nearly enough for a surprise attack with single warheads. But this advantage was counterbalanced, first, by our very large--and growing--advantage in warheads, since only we possessed MIRVs (Multiple Independently Targetable Re-Entry Vehicles) and, second, by our insurmountable superiority both in numbers and in the technology of long-range bombers, on which there was no limitation.

The SALT agreements of 1972 might well have achieved the objective of strategic stability. But both domestic and technical factors caused the accords to become increasingly controversial. The Viet Nam War and Watergate disintegrated the political consensus behind our defense and arms control policy just when technology was undermining its strategic premises. In the climate of collapsing confidence, groups usually associated with humane views came to advocate that the only way to keep our Government from using nuclear weapons was to deprive it of all alternatives to a strategy geared solely to the destruction of the Soviet population; never mind that the targeting of civilians guaranteed mutual annihilation. The other end of the spectrum disdained the proposition that we lived in a new world. It insisted that arms control was a trap and a delusion.

As we consumed ourselves in disputes over negotiations and weapons systems, SALT I, at first widely acclaimed, was drawn into that vortex. SALT II never emerged from it. Advocates of arms control belittled the extent to which it was being overtaken by technology. Opponents focused on the numerical "advantage" that SALT I allegedly gave the Soviets, overlooking that the agreed totals reflected exactly the level of U.S. forces the Pentagon chose long before there was any thought of arms control and that we retained a large numerical edge in warheads and airplanes.

Almost totally obscured in this debate was the reality that multiple warheads were making the traditional SALT approach obsolete. In SALT I a rough balance in the two sides' delivery vehicles substantially reduced the possibility of surprise attack. But multiple warheads--far exceeding the number of launchers--were bound to restore the advantage of the attacker, who could hope to overwhelm the opponent's fixed missile sites even with equal numbers of missiles and warheads on both sides. The side striking first would have an advantage--thus reviving the destabilizing danger of surprise attack. From this point of view, a "freeze" would perpetuate an inherently precarious state of affairs.

Fairness compels me to point out that the decision to proceed with MIRVs was taken by President Johnson and was made irrevocable in the Nixon Administration. We proceeded because in the climate of the Viet Nam period we were reluctant to give up the one strategic offensive program that was funded with which to counter the rapid Soviet missile force buildup; because we doubted that the Soviets could achieve accuracies to threaten our missile force in the foreseeable future; and because the Soviets ignored our hints to open the subject of a MIRV ban in the SALT talks. Whatever our reasons, there can be no doubt that the age of MIRVS has doomed the SALT approach.

The controversy over whether arms control was a boon or a trap--and some ill-considered comments on the feasibility of nuclear war--left defense policy increasingly at the mercy of the exploding public concern about the dangers of nuclear war. No democratic leader can govern any longer without demonstrating his devotion to peace. The Reagan Administration soon learned that the assault on what it called the "fatally flawed" SALT II treaty made for better campaign rhetoric than foreign policy. It compromised on the strange course of observing but not ratifying SALT II. The Administration has proclaimed its devotion to arms control, and I accept its sincerity. The challenge it faces is to resolve, finally, the intellectual problem of how to ensure strategic stability amid the revolution wrought by thousands of warheads on only hundreds of launchers.

The Present Dilemma There is a "flaw" in SALT II, though not the one usually discussed. It is that SALT limitations were expressed in terms of numbers of delivery vehicles at the precise moment when the increase in the accuracy and number of warheads caused numerical "equivalence" to be more and more beside the point. With each side possessing the capability (the Soviets' actual, ours latent) of making its opponent vulnerable, arms control after a decade of negotiations had returned to its starting point.

This problem cannot be solved simply by deep reductions in delivery vehicles. Given the disproportion between warheads and launchers, reductions either are irrelevant to the danger of surprise attack or, perversely, increase it. With present weapons, the greater the reductions, the fewer would be the targets for a first strike and the greater would be its calculability.

This is well illustrated by President Reagan's Eureka College speech of May 9, 1982, which contains the basic American proposal for the Strategic Arms Reduction Talks (START). It was an advance over an uncontrolled arms race because it set a ceiling. It was an advance over SALT in relating the ceiling to warheads rather than launchers. And it stressed significant mutual reductions of strategic forces. It was a brave first attempt that unfortunately did not solve the root issue of multiple warheads. Even were the Soviets to accept our proposal, the Eureka scheme would-at best maintain the existing balance; it would almost surely worsen rather than ease our dangers. A quick glance at the numbers involved illustrates the problem.

Under SALT n about 5,000 Soviet land-based warheads would be aimed at 1,054 American launchers--a ratio of less than 5 to 1. The Eureka proposal would reduce the permitted warheads to 2,500 on at most 400 launchers. Even were it technically feasible to distribute warheads in this manner (and the Soviets would have to redesign their entire strategic force to do so), this would give the side striking first an advantage in warheads to targets of better than 6 to 1. And at these lower numbers of launchers an attack would be far more calculable.

The Eureka proposal would also establish limits of 2,500 to 3,000 sea-based warheads. This would force a reduction of our submarine force from the 42 permitted under SALT to 15 or fewer (depending on the type). We could keep at most ten vessels at sea at one time, vs. the current 25 to 30. If there are any advances in antisubmarine warfare technology, as is probable, arms control will have increased the vulnerability of both our underwater and our land-based strategic forces--the supreme irony. The Soviet proposal of a flat 25% reduction in launchers, while simpler, suffers from the identical disability.

In short, a negotiation begun more than a decade ago to enhance stability and reduce vulnerability is in danger of achieving the opposite. Arms control is heading for an intellectual dead end.

WHERE DO WE GO FROM HERE?

We face two related tasks. First, arms control requires not so much a new proposal as a fresh concept. Second, it must become an organic part of defense policy. This requires that we return to first principles. The principal cause of instability with current weapons systems is the disproportion between warheads and launchers. All the remedies that have been tried are vulnerable to technology: hardening to accuracy, sea-based systems to advances in antisubmarine technology. There is no effective or intellectually adequate solution to this problem except to seek to eliminate multiple warheads within a fixed time, say ten years.

Fortunately technology, which creates the problem, can offer a solution. According to published literature, it is possible to develop a mobile missile that could be protected in a heavily armored canister. Its mobility alone could complicate the task of the attacker. Moreover, the new missile could--and should--be equipped with a single warhead. With strategic forces of such design, numerical limits would be both simple to establish and far more significant than under SALT n or START.

Once we decided on such an approach, we could proceed with it either as part of an arms control agreement or unilaterally as part of our defense policy. For example, we could propose to reduce and transform the strategic arsenals of both sides to a low number of single-warhead missiles over a ten-year period. The totals should be set at the lowest number that could be monitored; that is, at a level where a violation significant enough to overturn it could not be hidden. The permitted number of missiles may be as low as 500; at any rate, the number of warheads in this scheme would be only a small fraction of current totals, probably 20% or less of the Eureka scheme. Each side would be free to choose whether the permitted missiles would be mobile or in silos. Mobility would reduce the incentive of surprise attack, but equivalence at low numbers of single-warhead missiles would, in any event, assure considerable stability.

This course does not depend on Soviet agreement. It should be pursued whatever the Soviet reaction. If they refused our proposal--this one or another embodying the same concept--the U.S. could announce that after a certain date, say 1990 (or before then if the new missile could be developed earlier), it would deploy no more MIRVed land-based intercontinental missiles but would emphasize single-warhead launchers, the majority mobile. The size of that force would be geared to the number of warheads deployed by the Soviets; we would reserve the right to match each Soviet warhead with single-warhead missiles of our own. In practice, we would almost certainly choose a lower number that we calculate could survive the maximum Soviet attack capable of being launched. The purpose would be to increase the number of targets the Soviets would have to hit but without increasing our capacity for surprise attack. We would gradually phase out our MIRVed missiles. If the Soviets agreed to a formal proposal, schedules for the mutual destruction of MIRVs would be negotiated. If they refused, we would build up single-warhead missiles to a level consonant with our security. The Soviets could always put a ceiling on our deployment by cutting the number of their warheads.

This scheme should pose no insurmountable verification problems. Fixed launchers can be detected through national technical means; existing Soviet MIRVed iCBMs could not be made mobile; development of a new mobile MIRVed ICBM would require extensive testing, which could be detected and would therefore be proscribed. Mobile single-warhead missiles would be more difficult to detect; this is why agreed numbers would have to be sufficient so that they could be exceeded only by a violation our means of detection would not miss. Obviously, the more airtight the inspection, the smaller can be the numbers. Only missiles tested solely with single warheads would be permitted; any tested with a MIRV warhead would be proscribed.

No one can predict how the rigid Soviet bureaucracy would react to this approach. It may upset too many vested interests. The new leadership may be too dependent on military support to challenge its military-industrial complex. Yet sometimes an impasse can be broken only by a daring departure; surely the nitpicking SALT negotiations offer little hope for the traditional approach. If the Soviets can ever be interested in stability and in easing the economic burden of the arms race, they should--probably only on second thought--study this scheme with care. Like the Eureka approach, it requires them to redesign their forces; unlike the Eureka approach, it reduces their vulnerability. And upon reflection, the Soviets must realize that, one way or another, we will cure the vulnerability of our forces and in the process will almost surely enhance the vulnerability of theirs.

If the Soviets refuse to discuss such a proposal, one of three conclusions is inescapable: a) their arms program aims for strategic superiority if not by design then by momentum; b) they believe strategic edges can be translated into political advantages; c) arms control to the Soviets is an aspect of political warfare whose aim is not reciprocal stability but unilateral advantage.

Where does this leave the MX? A presidential commission is studying that question. I will address one issue: Should we have a "counterforce" capability (an ability to strike accurately at Soviet missile silos or command centers), or should we continue to aim for "assured destruction" of civilian and industrial targets? Ever since the Soviets began to approach strategic parity, it should have been obvious that a strategy aiming at civilian destruction was an irrational, suicidal, indeed nihilistic course that no President could implement. Undiscriminating slaughter is not a defense policy but a prelude to unilateral disarmament.

Similarly, why should a Soviet counterforce capability--as now exists--be treated as consistent with strategic stability, while our attempt, represented by the MX, to provide a much smaller means to respond is considered as somehow destabilizing? If the U.S., by its abdication, guarantees the invulnerability of Soviet missile forces while the Soviets keep ours exposed, any Soviet incentive for serious negotiation will vanish. A secure Soviet first-strike capability poses an unprecedented danger--ultimately that it may some day be used, in the near term that it may increase Soviet willingness to run risks in regional crises.

Whatever level of MX deployment is recommended by the Scowcroft Commission should be strategically meaningful beyond a mere token deployment. At the same time, the MX, like the new single-warhead missile, should be an organic part of an arms-control strategy. To this end, we should offer to postpone MX deployment if the Soviets agree to destroy MIRVed null (their heavy missiles) over three years starting in 1986, and to abandon MX altogether once the SS-18s are dismantled.

This analysis has been confined to land-based missiles. Were the Soviets to show interest in the scheme outlined here, account would have to be taken of sea-based forces. Just as we cannot be asked to ratify our own vulnerability in land-based forces, the Soviets should not be expected to acquiesce in U.S. submarine-launched missiles capable of surprise attack. Specifically, as part of the agreement proposed here, we should be prepared to move to single warheads at sea as well, though over a longer period, say 15 years, because of the long lead times. In that case, the submarines would have to be made smaller and less expensive. It would be too risky to put so many eggs in one basket, as is the case with the current Trident submarines, each of which carries 24 missiles. A new regimen would be required as well for heavy bombers.

The deployment-arms control scheme would then look as follows:

A. The U.S. would make a fundamental decision to shift to single-warhead missiles as soon as possible. Ideally, this decision would be reflected in an agreed ceiling at a very low number--perhaps 500--negotiated with the Soviets. An agreement should also limit throw weight to prevent development of huge single-warhead weapons.

B. If the Soviets refused such a scheme, we would proceed unilaterally toward our goal. The final size of a single-warhead force would depend on the number of warheads in the Soviet force and on what we need to assure our invulnerability.

C. The U.S. would begin deploying MXs starting in 1986. It would be prepared to postpone deployment if, before 1986, the Soviet Union agreed to a schedule by which its SS-18s would be destroyed over a three-year period starting in 1986.

D. Both sides would also agree to dismantle the remaining land-based MIRV forces starting in 1990.

E. In that case, both sides would agree not to increase the number of warheads on MIRVs while they remain in the force.

F. Other mixes are possible. For example, a small number of MIRVed missiles and bombers, no more than 200, could be joined with a reduced single-warhead deployment, say 300.

This approach would be a serious test of Soviet intentions. It would conclusively end the danger of a first strike. It would establish clear equivalence. It would transcend the SALT and START debate and put strategy and arms control in a coherent context. If refused, it would be a clear signal of a Soviet bid for superiority; we would draw the appropriate conclusions. If we proceeded unilaterally, nevertheless, it would be a major contribution to strategic stability and U.S. security.

Of course, even the achievement of strategic stability would open up areas of concern now dormant. It would bring to the fore the pressing need to build up conventional forces to deter non-nuclear challenges. That problem would be addressed in a new environment. For all parties would know that they have taken--at last--a big step toward avoiding nuclear catastrophe. This is an imperative that humanity demands and reality imposes. - This file is automatically generated by a robot program, so viewer discretion is required.